SDRAM performance bottleneck

Hi guys, 

I am interfacing a 16MB sdram with OMAPL137. I have initialized the sdram as per the datasheet specifications and now want to run a profiling test to see the performance. When i run a simple loop (as below), the performance looks really poor (as compared to that of a SRAM). 

My profiling loop. 

loopOut = 1; loopIn = 1500; //instruction reads a 8bit sdram location and writes it on another location in sdram

for (loopOut = 0; loopOut < LOOP_OUT_MAX; loopOut++)
{
   for (loopIn = 0; loopIn < LOOP_IN_MAX; loopIn++)
   {
      *(pDst + loopOut)= * (pSrc + loopOut);
   }
}

My code is running from SDRAM as well. It takes roughly 5ms to complete this loop. My DSP is clocked at 160Mhz and sdram is clocked at 80Mhz. Any suggestions to improve on these timings ?

Thanks, 

AQ